Gnocchi: slow starch and an empty composition record

Across 19 randomised controlled trials, resistant starch lowered fasting plasma glucose by 0.09 mmol/l against digestible starch (95% CI -0.13 to -0.04), and by 0.16 mmol/l once intake passed 28 g a day or the trial ran beyond 8 weeks. Those trials fed supplemental resistant starch rather than gnocchi, and the effect on fasting insulin did not reach significance. The distinction matters, because the only human trial that fed the dumplings themselves is smaller and much older. Seventeen adults with type 2 diabetes ate 165 g of potato dumplings against an equal-carbohydrate portion of white bread. Plasma glucose ran lower after the dumplings at 90, 120 and 150 minutes (P < 0.05, P < 0.01 and P < 0.05), and the insulin response did not differ. Those were 17 patients eating a single meal in 2001, and the dumplings were an Italian commercial product that was not standardised. The laboratory arm of that same trial hydrolysed about 30 percent less starch from potato dumplings than from pizza or bread over 5 hours, and electron microscopy showed a compact structure that keeps enzymes away from the starch. That arm found no difference in resistant starch content, sugar diffusion or digesta viscosity, and it measured test foods in glassware rather than people.

What your body absorbs

• In the laboratory arm of the same trial, only about 70 percent as much starch was digested from potato dumplings as from pizza or bread over 5 hours, and electron microscopy showed a compact structure that keeps enzymes away from the starch. — test foods, not people: in vitro alpha-amylase hydrolysis of the same four Italian wheat-based foods (n = 4 test foods) Evidence D source
  Starch hydrolysed after 5 h about 30% lower for potato dumplings than for the other foods, with no difference in resistant starch content, sugar diffusion or digesta viscosity

What cooking and storage change

• Across 19 randomised controlled trials, resistant starch lowered fasting plasma glucose by 0.09 mmol/l (95% CI -0.13 to -0.04) against digestible starch, and was larger, at 0.16 mmol/l, above 28 g a day or beyond 8 weeks. — adults in 19 randomised controlled trials, resistant starch against digestible starch (n = 19 trials) Evidence A source
  Fasting plasma glucose effect size -0.09 mmol/l (95% CI -0.13 to -0.04), -0.16 mmol/l above 28 g/day, HOMA insulin resistance -0.33 (95% CI -0.51 to -0.14)
• In 30 women with raised fasting glucose, 250 g of potato baked then chilled cut insulin by about 26 percent at 15 minutes and GIP by about 41 percent, compared with the same weight boiled and eaten hot. — 30 women, mean BMI 32.8, fasting glucose 110.5 mg/dL, two test meals of 250 g russet potato (n = 30) Evidence B source
  At 15 and 30 min glucose down 4.8% and 9.2%, insulin down 25.8% and 22.6%, GIP down 41.1% and 37.6%, insulin and GIP area under the curve lower
• When hard-cooked potato replaced one third of the carbohydrate in a rice meal, peak glucose and insulin fell, but the benefit was gone by the next meal even though the combined 540-minute glucose curve stayed lower. — healthy adults, acute test meals of rice with hard-cooked or soft-cooked potato, measurements over 540 min Evidence B source
  Hard-cooked potato with rice lowered glucose iAUC, peak glucose, glycemic excursion, insulin iAUC and peak insulin; soft-cooked raised glycaemic variability

Safety, limits and interactions

• Chilled packaged gnocchi held at the label temperature of 8 degrees Celsius or at 12 degrees produced no Clostridium botulinum growth or toxin in 75 days, and the safety margin depended on 0.09 percent sorbic acid once the pack reached 20 degrees. — eight commercial gnocchi formulations in challenge tests, plus attempts to isolate C. botulinum from gnocchi and its ingredients (n = 8 formulations) Evidence D source
  No growth or toxin in 75 days at 8 and 12 degrees C, protection factor above 4.2 for proteolytic and above 6.2 for non-proteolytic C. botulinum, under 25 spores/kg estimated in gnocchi
• Bacillus cereus was present in 33 percent of chilled gnocchi samples, below 100 CFU per gram while refrigerated at 8 degrees Celsius, and counts rose once storage was warmer or longer in recipes made without sorbic acid. — retail and laboratory gnocchi samples under varying storage conditions, plus spore challenge tests at 8, 12 and 20 degrees C Evidence D source
  B. cereus in 33% of samples, under 10(2) CFU/g unstored and at 8 degrees C, growth inhibited by sorbic acid with citric or lactic acid to pH 5.0, lactic acid alone failing at 12 and 20 degrees C

What people believe that the data does not support

• Across 103 observational studies, total potato intake carried no higher risk of hypertension or cardiovascular disease, while non-fried potato intake carried 5 percent higher diabetes risk (RR 1.05, 95% CI 1.01 to 1.10). — adults in 51 articles covering 103 observational studies (n = 103 studies) Evidence C source
  Non-fried potato and diabetes RR 1.05 (95% CI 1.01 to 1.10), fried potato and diabetes RR 1.16 (1.04 to 1.30), total potato and CVD RR 0.97 (0.91 to 1.03)

Who this works differently for

• In 17 adults with type 2 diabetes, 165 g of potato dumplings gave a lower plasma glucose response than an equal-carbohydrate portion of white bread from 90 to 150 minutes after the meal. — 17 adults with type 2 diabetes, 11 men and 6 women, each food giving 50 g available carbohydrate, glucose followed for 180 min (n = 17) Evidence B source
  Plasma glucose lower than white bread at 90 min (P < 0.05), 120 min (P < 0.01) and 150 min (P < 0.05), insulin response no different

The bottom line

Cooking and cooling do measurable things to potato starch, and both trials that show it used potato rather than dumplings. In 30 women with raised fasting glucose and a mean BMI of 32.8, 250 g of russet potato baked and then chilled cut insulin by 25.8 percent at 15 minutes and GIP by 41.1 percent, against the same weight boiled and eaten hot. Those were two acute single meals in women only, and satiety did not change. In a second acute trial, hard-cooked potato replacing one third of the carbohydrate in a rice meal lowered peak glucose and peak insulin. The benefit had gone by the next meal, even though the combined 540-minute glucose curve stayed lower, and that abstract does not say how many people ate the meals. The potato reputation is worse than the pooled data. Across 103 observational studies in 51 articles, total potato intake carried no higher risk of cardiovascular disease (RR 0.97, 95% CI 0.91 to 1.03) and none of hypertension. Non-fried potato intake carried 5 percent higher diabetes risk (1.05, 1.01 to 1.10) and fried potato 16 percent (1.16, 1.04 to 1.30). Intake in all of that was self-reported, and boiled dumpling dough does not fall cleanly into either the fried or the non-fried column. Who should be careful, and here the honest answer is that the safety material on this page is food microbiology of the chilled supermarket pack. Eight commercial gnocchi formulations held at the label temperature of 8 degrees Celsius, or at 12 degrees, produced no Clostridium botulinum growth or toxin in 75 days. That was a laboratory inoculation study of Italian products from 2004, and its whole margin at 20 degrees depended on 0.09 percent sorbic acid being in the recipe. Bacillus cereus turned up in 33 percent of gnocchi samples, below 100 CFU per gram unstored and at 8 degrees, with counts rising in warmer or longer storage and in recipes without sorbic acid. Prevalence in a 2001 laboratory survey is not illness, and none of that work measured people. Keep the pack cold and inside its date. One thing belongs to the dough rather than the pack: the potato dumplings fed in the single human trial here were counted with pizza and bread among wheat-based foods, so a commercial pack is not something to assume is gluten-free. We found no study of this food and allergy, pregnancy, children or medicines, and that silence is not a clearance. Neither is the composition. Gnocchi has no USDA record and no second independent composition number, and nobody has measured its own glycaemic index or what boiling and then chilling it does to its resistant starch.

Sources

1. gnoc-r5-02 · in vitro starch hydrolysis nested in a randomised trial
   The percentage of starch hydrolysed after 5 h in vitro hydrolysis with alpha-amylase was about 30 % lower for potato dumplings than for the other foods.
   Br J Nutr, 2001 · checked 2026-09-24
2. gnoc-r5-03 · systematic review and meta-analysis of 19 randomised controlled trials
   The effect of resistant starch on fasting plasma glucose was significant (effect size (ES) -0·09 (95 % CI -0·13, -0·04) mmol/l, P = 0·001) compared with digestible starch.
   Br J Nutr, 2021 · checked 2026-09-24
3. gnoc-r5-04 · randomised crossover trial in females with elevated fasting glucose
   Results showed that the chilled potato elicited significant reductions at 15 and 30 min in glucose (4.8% and 9.2%), insulin (25.8% and 22.6%), and glucose-dependent insulinotropic peptide (GIP) (41.1% and 37.6%), respectively.
   Nutrients, 2019 · checked 2026-09-24
4. gnoc-r5-05 · randomised controlled acute trial in humans
   HP + R significantly reduced postprandial glucose iAUC, peak glucose and glycemic excursion...The hypoglycemic effect did not extend to the second meal, though composite iAUCglucose over 540 min was reduced.
   Nutrients, 2026 · checked 2026-09-24
5. gnoc-r5-07 · laboratory challenge test study on gnocchi formulations
   For all gnocchi stored at 8 degrees C (as recommended by the manufacturer) or 12 degrees C (mild temperature abuse), growth and toxin production were not detected in 75 days.
   Int J Food Microbiol, 2004 · checked 2026-09-24
6. gnoc-r5-08 · in vitro challenge test and laboratory analysis of retail gnocchi
   A survey of gnocchi under varying storage conditions, revealed that, although B. cereus was found in 33% of the samples, the contamination level was lower than 10(2) CFU/g for the unstored and the refrigerated (8 degrees C) samples.
   Int J Food Microbiol, 2001 · checked 2026-09-24
7. gnoc-r5-06 · systematic review and meta-analysis of observational studies
   nonfried potato intake was significantly associated with increased diabetes risk (1.05; 95% CI: 1.01, 1.10) but not hypertension (1.06; 95% CI: 0.97, 1.15)
   Nutr Rev, 2025 · checked 2026-09-24
8. gnoc-r5-01 · randomised controlled crossover trial in patients with type 2 diabetes
   Plasma glucose response, measured for 180 min, was significantly lower after the potato dumplings than after white bread at 90 (P < 0.05), 120 (P < 0.01), and 150 (P < 0.05) min.
   Br J Nutr, 2001 · checked 2026-09-24